Pipelines are used to efficiently transport fluid commodities from one location to another and, generally, span long distances. Pipelines are typically monitored and controlled to ensure the integrity of the pipeline. This is normally accomplished from a central control center where equipment settings and measured parameters are monitored and controlled. Leaks can be detected by measuring various parameters, particularly flow rate and pressure. This monitoring depends on calculating the mass of the contents of the pipe (fluid), and observing over a period of time whether the contents of the pipe changes in a manner indicative of fluid leaving the pipe at an unmeasured location, perhaps through a rupture in the pipeline. The effect of a leak on the measured parameters and thus on the calculated mass of fluid within the pipe is mathematically relatable to the mass of fluid which is leaving the pipe through the rupture. Using existing methods, large ruptures can be detected in relatively short periods of time. Small ruptures, however, if at all detectable, require longer periods of time to accumulate the necessary volumetric discrepancies to indicate excessive system measurement imbalance, which may result in a greater amount of fluid leaking from the system into the surrounding environment. These systems depend on a balancing algorithm to determine if fluid is leaving the pipeline at an un-metered location to determine if there is a rupture. As may be appreciated, these prior art systems require that a sufficient volume of fluid has left the pipeline through the rupture before detection of the rupture can occur. This can result in a significant environmental impact in the area surrounding the rupture as inferential measurement systems are commonplace and relatively lethargic with respect to hydraulic wavespeeds and the ensuing segmental decompressions that follow rupture events in closed pressurized hydraulic networks.
Calculating the mass of the contents in the pipe is subject to many inaccuracies (e.g., errors in measurement of the mass of the fluid entering and leaving the pipe, inaccurate knowledge of the changes in the physical space within the pipe due to temperature fluctuations in the pipeline walls and the minimal knowledge of the actual temperatures and pressures within the pipeline, transducer calibration and/or configuration correctness).
The approximate location of a piping rupture can be determined using existing methods for large ruptures by examining the effects of the rupture on the flow rate and pressure at locations where measurements are available. Locating the rupture under existing methods is subject to several possible inaccuracies related to measurement system and ancillary instrumentation system parameterization, configuration, and calibration. As the volume of fluid exiting through the rupture drops to a small fraction of the flow rate of the pipeline, the effect of the missing fluid will drop below the accuracy with which the parameters used to calculate the location are measured. Additionally, common inferential mass metering systems often operate within nonlinear regions of equipment operation thus degrading leak detecting capabilities.
There is a need for a system and method to rapidly detect and accurately predict the location of a destabilizing event in a pipeline network. Others have attempted to develop systems that are more effective in identifying ruptures than those described above, but these systems only identify leaks and ruptures and not destabilizing events. These destabilizing events may not be a leak or rupture, but impact the operation of the pipeline. These systems, however, compare sensed conditions with pre-identified stored leak profiles. These systems do not identify unstable operations that are not leaks or ruptures, the locations of these unstable operations or the source(s) of concern. These systems are not dynamic, as such, these systems do not adequately respond to abnormal changes in the operation of a pipeline network that are not indicative of the presence of a leak or a rupture.
U.S. Pat. No. 5,361,622 to Wall discloses a device and method for the detection of leaks in a pipeline. Wall utilizes a transducer for measuring instantaneous pipeline pressure. Wall utilizes a computer to compare rates of change of the measured pressure at successive preselected and fixed timed intervals at isolated nodes. The computer compares the measured rate of change with a preselected total change in a specified characteristic at that node. When the measured rate of change exceeds the preselected maximum total change, an alarm is triggered indicating the possibility of a leak or rupture. These maximums, however, do not vary or adjust for variations in the pipeline operation. Wall does not aid in the determining the location of the leak or rupture because its analysis is site specific. Wall provides for local or isolated analysis at select points within the pipeline. Wall does not corroborate these changes with any changes occurring at other nodes. This lack of corroboration leads to a greater potential for false alarms.
U.S. Pat. No. 5,388,445 to Walters et al discloses a method and an apparatus for detecting a wave front caused by the onset of leaks or other transient events in a pipeline. Walters seeks to detect a significant wave front traveling in the high noise environment characteristic of pipelines, and to accurately measure its amplitude and time of arrival. Walters discloses detecting a wave front indicative of a transient event occurring in a pipeline by measuring a characteristic related to pressure of a fluid in the pipeline with a measuring device positioned at a given point on the pipeline, and outputting analog signals proportional to the pressure. The detection of the wave front arrival time may be used directly to sound an alarm. The amplitude may be used to find the location and/or size of the leak. Walters, however, does not provide means for discounting (i) external influences on the pipeline, which may produce wave fronts but are not leaks or (ii) internal influences such as changes in the viscosity or strip operations within the pipeline, which again may produce wave fronts but are not leaks. The model employed in Walters is a fixed model. It does not adjust or learn based upon varying pipeline operations and external influences; rather, it relies upon a fixed reference event to determine leaks. Accordingly, the system disclosed by Walters may be prone to false leak indications. Furthermore, Walters looks at measurements obtained during affixed time window and does not account for events that may start during one window and end during another window. Accordingly, the system may fail to detect a leak nor does it attempt to identify or diagnose destabilizing events that are not leaks but that may adversely affect the pipeline network's efficiency of operation or safety.
U.S. Pat. No. 5,428,989 to Jerde et al discloses a method and an apparatus for detecting and characterizing a leak using a pressure transient. Jerde provides a plurality of pressure monitoring stations spaced along a pipeline at known distances from each other. The stations generate an arrival time signal when a pressure wave front is detected. The system can determine whether the arrival time signals from the plurality of monitoring stations correspond to the same event.
U.S. Pat. Nos. 6,389,881 and 6,668,619 both to Yang et al disclose an acoustic based method and apparatus for detecting and locating leaks in a pipeline. Yang utilizes a system for detecting leaks using the acoustic signal generated from a leak event in a pipeline. The system requires the use of non-industry standard specialized components mounted to the pipeline at selected locations. A pattern match filter is used to reduce false alarm rate and improve leak location accuracy by comparing acoustic waves generated by a leak with stored previously recorded signature leak profiles. Yang utilizes a time stamp of the acoustic signal to pinpoint the location of the leak. The system listens for acoustic waves to determine whether or not a leak is present. Normal operating events (such as pump start ups, etc.) and isolated events (such as a piece of equipment hitting the pipeline) may generate an acoustic signal that travels through the pipeline but not a pressure wave characteristic of a leak. As such, the Yang system is more prone to issuing false alarms.
U.S. Pat. No. 6,970,808 to Abhulimen et al discloses a method for detecting and locating leaks in a pipeline network in real-time utilizing a flow model that characterizes flow behavior for at least one of steady and unsteady states, which corresponding to an absence and a presence of model leaks in the pipeline network. A deterministic model is provided to evaluate at least one of a leak status and a no leak status relating to the pipeline network using deterministic criteria.
These prior art systems are not proactive systems; rather, these systems are reactive in nature relying upon comparisons to historical data or predetermined modeled results in order to determine whether or not a leak is present. These prior art systems essentially reset to the original threshold after each detection of a potential leak. The thresholds of prior art do not rescale themselves in real time to take into account normal changes in pipeline operation. The thresholds are generally high, which overlook changes due to small leaks. These prior art systems do not rely upon present pipeline activity in order to determine the presence of a leak or rupture, which may require further action nor do they perform corroborating steps to confirm a destabilized event, nor do they diagnose the source of the destabilizing event (e.g., process error, equipment malfunction). These systems do not continuously adjust detection thresholds to account for the various normal operating modes of pipeline networks. While it is desirable to maintain a stable pressure and stable flow through the pipeline network, changes in pressure and flow are common due to changes in viscosity of the various fluids flowing through the network, equipment malfunction, etc. These prior art systems are not capable of accounting for these variations in pipeline activity. As such, these prior art systems may generate a significant number of alarms indicating a potential leak. Each of these alarms requires some form of action by a pipeline operator. There is a need for a dynamic logic-based system that accurately responds to the varying operating conditions present in a pipeline network while accurately identifying destabilizing events including leaks and ruptures while minimizing false alarm events. Furthermore, there is a need for a system that effectively manages the generation of alarms.